So what your saying is that in order to have a functional system you need more than 2gb of hard drive space...
Only 48 mb of ram? My limited experience with X has taught me that in order to run any software of consequence (i.e.: StarOffice, Netscape, etc.) I really, really want 128mb + a good fast cpu.
My point is only that in order to run generic software under X the base system recommendation now is probably not dissimilar than the W2K/WXP ones.
So... I take it, mangle it, change it... then I sell the binary.
The source is freely available to anyone who wants to pay for it -- it is in a nice leather bound dead tree (paper only, no media). Total shipping weight 50lbs or so and total cost in the range of 30k or so (very nice binding, only the best paper, hand packed).
Yet we run several hundred users a day on configs like that (60 to 80 per box).
That is for MS Office and several vertical market applications (some of which take 45 to 60mb per user). Remember, the claim to fame of WTS/W2K with TS is that it will share the same copy of code where it can. A typical snapshot of a box with ~60 users on it is 25% cpu utilization (quad cpu box on this one), 800mb free RAM and low if any disk utilization.
That's 60+ active users as well. Our original estimates were that these boxes would go to 80 users.. and they can.
A quick peruse of the server requirements seems to indicate that this will be as server intensive (memory) or more so than a typical NT Terminal Server / Windows 2000 Terminal Server implementation. CPU wasn't mentioned too much but memory (2gb for 20+ users) is the same amount we use as a ballpark for WTS/Metaframe with 60 to 80 users. We do generally go dual CPU (or quad if very cpu intensive).
Does anyone have any sizing info on X in a multi-user environment?
License management: this is an issue. It shoudl be noted that the admin's should also know the version of installed OS for any OS on every box to begin with. This issue is only partially mitigated by not going the license route. In order to adequately maintain the environment you need to know what the environment is.
Forced upgrades: can you say library upgrades? I'm willing to bet you will have forced upgrades in the Linux world based on new applications within the next 12 months let alone 5 years (i.e.: 5 years from release of W95 to roughly now). This cost is included in a decent ROI as either maintenance or upgrade costs.
20$ per user as quoted was $20 per user per month -- i.e.: it is a repeating expense. It is significant but when distilled to that number most bean counters add it automatically to their "head" price and get on with life.
Downtime losses -- if you want to be in business then buy the appropriate hardware and software. Windows 98/95/ME are not appropriate where downtime is a concern. Not an argument -- just a statement of fact. 2000 and NT are much more appropriate and Linux has stability that is in at least the same ballpark. The argument is one of what are you giving up to go Linux over NT/2000?
New hardware is an issue. Think you can run the latest Star Office and get productivity on that P133? On a P200 with 64mb of RAM NT seems to hold neck and neck for office apps with Linux. Add another 64mb of RAM and NT may even beat it. Frankly, whenever you go graphical on either platform you spend hardware bucks. If we are talking non-graphical Linux wins. Do you really mean that you are going to have those users not using X? If so the argument shifts dramatically and a P133 is adequate for Linux. Of course Linux looses immediately due to a lack of user interest -- sad but true.
...is your friend. No user owns their own workstation. No user is ever root (or admin).
We have been restrictive but have opened up over time to make sure that essential functions are available (i.e.: task manager can be quite handy if they have an out-of-control app so they can kill it).
This is a managed implementation. That means that the SA's and support people working in concert with the business leaders had to lay out a road-map of what was allowed and what wasn't. Then they had to implement it.
No user installed games, no user installed screen savers, no user installed software. Yep, it's Draconian. It is a business choice implemented through technology.
And understand that what you have forgotten is that MS has covered a large part of the licensing field -- if you want to use their products or most products developed using their products (support libraries) you need to license the MS OS.
Say what you will about MS Marketing but they are not dumb or short sighted.
..I've seen it. They committed the infraction. They paid. US $129,000 is chicken feed. It would have been less if they had tighter management of their software.
The savings will add up -- and will be less than $20 per user per month assuming 0 negative impact for making the choice. Any negative impact at all from the choice can easily cost millions of dollars (10's of millions of pesos).
I'm not saying don't do it. Do it -- just make sure that it is clearly thought out and that the city is protecting the public trust when they do.
Which we have seen less than 5 occurrences in the last year (200 users at location A, 40 hour work week, less than 5 blue screens). In each case it was a hardware issue (memory/motherboard/etc.). On servers we have not seen it at all.
I'm not saying you won't see it but I am saying that if you have decent IT policies and procedures you will not see it often.
..but the justification is not there. Business users generally want to do their job and care little about the OS involved.
It's not a matter of being brave. It is a matter of having covered any gaps in the offering. Of the 200 users installed above 0 could exist in a Linux environment. They all use 2 custom Windows applications that are specific to the health care industry. The vendors do not and will not support them in a different environment and that includes emulation.
As stated before Linux needs a killer app. An app to drive the purchase of workstations. The app's in question here justified the purchase of the infrastructure and had a defined pay-back time frame. These are things that make bean-counters happy -- vendor support and defined pay-back cycles.
Seems like the vertical market is prime for Linux... opportunity is knocking. Of the dozen or two major vertical market shop's I deal with none (0, zero, nada) are planning any desktop implementation of their software in a Linux environment. They may consider limited server support in a second tier environment someday in the future... but not yet.
This is important because the users I deal with are driven by the "killer" apps in their field -- they will buy the hardware and software package that does the job for them. The justification is in the pay-back.
Turning that around and putting it in the corporate IT position it is difficult for any IT department to justify a desktop rollout of Linux if neither the industry standard desktop software or the "killer" apps that the business needs to survive are native or supported on it. Justification has to be in dollars and cents. Justification is not just license costs but includes support costs. The killer apps are not driving Linux implementation yet. The killer apps are preventing Linux implementation as they are still requiring Windows. It's a double edged sword.
...does not learn to take criticism either externally or internally we, Linux users, will never reach a level of maturity or acceptance.
When Linux can be said to "run all the apps users want to run" then WE will have something to be even more proud of.
I guess I'm guilty of the unforgivable sin, running a MS product in a business environment and having it work. Didn't say it was perfect, just said it has been cost-effective.
My dream is that some day the same can be said for Linux as a generic desk top replacement. A useful measurement toward that goal would be an objective study showing the savings brought by using Linux in the work place. I'd even take a gander at a non-objective study. I'm looking for support metrics. I'm looking for gap analysis. I'm looking for total cost of ownership.
So far all I'm getting here has been a knee-jerk statement that I must be a MS weenie. I have used MS software. I have used Linux. I have used Mac OS 6.5/7/8/9. I cut my teeth on an APF Imagination machine. I used a Commodore 64 and a Commodore 128. I am using W2K to write this on.
I am probably more in the average of the technical population than anyone who is a software bigot -- the attitude of only one OS should be out there is troubling. The attitude that their is not room for any one specific OS is disturbing.
A knee-jerk reaction to any product because it is made by MS is on par with a knee-jerk reaction to any student who dresses in a dark trench coat. I expect more of professionals.
... the point is that the 10% of the problem that is not resolved if it is NOT planned for will soak up the savings.
Are their any systems designers/planners out there? Anyone who has done any coding on a large scale product? Anyone else who's seen what happens when 90% of the problem is addressed leaving a critical 10% to hang in the wind?
I am not condemning Linux on the desktop -- just pointing out that the product has not been proven to meet general desktop requirements. I would suggest that instead of a knee-jerk reaction of "well you don't know..." you might examine the shortcomings of the product base and honestly attempt to shore it up.
First -- if they are saving 10's of millions of pesos the maximum they are saving is in the neighborhood of 9.6 million dollars, assuming tens of millions of pesos is really 100 million. Simple math. You should be able to handle it.
About the limited savings I mentioned -- my point is exactly as stated above -- as you stated it. I have seen no reliable studies by any groups supporting any conjecture that a Linux install is cheaper in any way shape or form. What I have seen is industry standard support costs for Windows environments. It is easy to identify within those industry standard documents the support costs of the licensing of the software. That cost is close to what I represented. I'd suggest you look up Gartner to start with.
If you know of any study by someone with authority please do tell. After all your unsupported rebuttal was pretty much a troll.
Of the thousands of applications available today in vertical markets and of the thousands of applications being developed what percentage are being developed or available on the platform used by 90% of the OS users out there today? Windows is king on the desktop still. A fact that is still a fact no matter how you may try to dispute it (industry surveys of late).
I have worked for state and local agencies and governments in the US only. I can tell you from experience that the majority of programs they have to use and need implemented are windows based. Is this a cart or horse problem? Good question -- the issue/point here is that the bean counters need to make sure they know the answer and most times DO NOT.
The only stupid question in a system integration is the unasked one and most times bean counters need competent IT support to make those decisions. If the bean counters have not heard the reason that the objections I raised are invalid in their configuration then they need to ask more questions.
Mexico City is part of a sovereign nation. If you haven't noticed many nations, Mexico included, have some sort of intellectual property arrangement with the world at large. Intellectual property is protected in some manner.
X and Terminal Server/Citrix server differing but similar purposes. One does not replace the other. Perhaps if you were a little more agnostic you might consider that Windows has a spot in IT today. Terminal Server/Citrix enhances that spot. There is some overlap between X and Terminal Server/Citrix. Overlap that might help to drive competition and enhance the product.
The down side to this is that MS understands licensing. They understand Corporate and Government (national and international) IT. They understand MARKETING. The fact is that a Windows machine can interact in a mixed mode environment at little incremantal cost. A Linux machine will incurr licensing fees to use Windows based applications. It is the way the license is written. Advange: Windows.
I installed NT 4 on 200 workstations 2.5 years ago. We have updated service packs but not re-installed the OS on any machine. I don't get the point of the Redhat comment. Since that time we have installed NT and W2K on hundreds more machines. It is stable. It works today. It runs all the apps the users want to run.
Star Office is a great effort and it is coming of age. Other than having little faith in Sun.... (they are a company...they are owned by stockholders and are beholden to those stockholders... if the management is not making a profit they are out on their rears and the new management can orphan or attempt to modify the licensing of the product.) my only issue is that it is a second class citizen in the Office space. It is not the leader, the innovator or the standard. MS Office is.
Frankly, in this troll you have posted, I see no data and no solid argument. Got any facts? You may find I'm very receptive to them.
Slashdot Mirror
The mission of every public company in the US is incredible shareholder wealth. If you believe otherwise you are wrong.
Acutal clocked down has been as low as 266.5kbps and as high as 2mbps+.
Actual up has been.. sufficient...
Realize that speed tests are also somewhat OS specific. The same speed test on a W98 box is much worse... on 2000 or Linux very close.
Only 48 mb of ram? My limited experience with X has taught me that in order to run any software of consequence (i.e.: StarOffice, Netscape, etc.) I really, really want 128mb + a good fast cpu.
My point is only that in order to run generic software under X the base system recommendation now is probably not dissimilar than the W2K/WXP ones.
Most Linux installs require at least that if you intend on going graphical (X) and using any applications.
The most common format for data transfer today is paper and ink.
It may violate the spirit of the agreement but not the terms of it.
The source is freely available to anyone who wants to pay for it -- it is in a nice leather bound dead tree (paper only, no media). Total shipping weight 50lbs or so and total cost in the range of 30k or so (very nice binding, only the best paper, hand packed).
until one member of management changes. That's all it takes now-adays.
Yet the first demo's I saw of OSX were on Intel hardware about 2 years or so ago....
You need a killer app for them.
That is for MS Office and several vertical market applications (some of which take 45 to 60mb per user). Remember, the claim to fame of WTS/W2K with TS is that it will share the same copy of code where it can. A typical snapshot of a box with ~60 users on it is 25% cpu utilization (quad cpu box on this one), 800mb free RAM and low if any disk utilization.
That's 60+ active users as well. Our original estimates were that these boxes would go to 80 users.. and they can.
Does anyone have any sizing info on X in a multi-user environment?
What was the revolutionary part here?
Forced upgrades: can you say library upgrades? I'm willing to bet you will have forced upgrades in the Linux world based on new applications within the next 12 months let alone 5 years (i.e.: 5 years from release of W95 to roughly now). This cost is included in a decent ROI as either maintenance or upgrade costs.
20$ per user as quoted was $20 per user per month -- i.e.: it is a repeating expense. It is significant but when distilled to that number most bean counters add it automatically to their "head" price and get on with life.
Downtime losses -- if you want to be in business then buy the appropriate hardware and software. Windows 98/95/ME are not appropriate where downtime is a concern. Not an argument -- just a statement of fact. 2000 and NT are much more appropriate and Linux has stability that is in at least the same ballpark. The argument is one of what are you giving up to go Linux over NT/2000?
New hardware is an issue. Think you can run the latest Star Office and get productivity on that P133? On a P200 with 64mb of RAM NT seems to hold neck and neck for office apps with Linux. Add another 64mb of RAM and NT may even beat it. Frankly, whenever you go graphical on either platform you spend hardware bucks. If we are talking non-graphical Linux wins. Do you really mean that you are going to have those users not using X? If so the argument shifts dramatically and a P133 is adequate for Linux. Of course Linux looses immediately due to a lack of user interest -- sad but true.
We have been restrictive but have opened up over time to make sure that essential functions are available (i.e.: task manager can be quite handy if they have an out-of-control app so they can kill it).
This is a managed implementation. That means that the SA's and support people working in concert with the business leaders had to lay out a road-map of what was allowed and what wasn't. Then they had to implement it.
No user installed games, no user installed screen savers, no user installed software. Yep, it's Draconian. It is a business choice implemented through technology.
...without reservation.
Say what you will about MS Marketing but they are not dumb or short sighted.
The savings will add up -- and will be less than $20 per user per month assuming 0 negative impact for making the choice. Any negative impact at all from the choice can easily cost millions of dollars (10's of millions of pesos).
I'm not saying don't do it. Do it -- just make sure that it is clearly thought out and that the city is protecting the public trust when they do.
... /. is really alt.I.want.brown.nose.RS ... or was that alt.I.fear.because.I.know.Im.inadequate
I'm not saying you won't see it but I am saying that if you have decent IT policies and procedures you will not see it often.
It's not a matter of being brave. It is a matter of having covered any gaps in the offering. Of the 200 users installed above 0 could exist in a Linux environment. They all use 2 custom Windows applications that are specific to the health care industry. The vendors do not and will not support them in a different environment and that includes emulation.
As stated before Linux needs a killer app. An app to drive the purchase of workstations. The app's in question here justified the purchase of the infrastructure and had a defined pay-back time frame. These are things that make bean-counters happy -- vendor support and defined pay-back cycles.
Seems like the vertical market is prime for Linux... opportunity is knocking. Of the dozen or two major vertical market shop's I deal with none (0, zero, nada) are planning any desktop implementation of their software in a Linux environment. They may consider limited server support in a second tier environment someday in the future... but not yet.
This is important because the users I deal with are driven by the "killer" apps in their field -- they will buy the hardware and software package that does the job for them. The justification is in the pay-back.
Turning that around and putting it in the corporate IT position it is difficult for any IT department to justify a desktop rollout of Linux if neither the industry standard desktop software or the "killer" apps that the business needs to survive are native or supported on it. Justification has to be in dollars and cents. Justification is not just license costs but includes support costs. The killer apps are not driving Linux implementation yet. The killer apps are preventing Linux implementation as they are still requiring Windows. It's a double edged sword.
When Linux can be said to "run all the apps users want to run" then WE will have something to be even more proud of.
I guess I'm guilty of the unforgivable sin, running a MS product in a business environment and having it work. Didn't say it was perfect, just said it has been cost-effective.
My dream is that some day the same can be said for Linux as a generic desk top replacement. A useful measurement toward that goal would be an objective study showing the savings brought by using Linux in the work place. I'd even take a gander at a non-objective study. I'm looking for support metrics. I'm looking for gap analysis. I'm looking for total cost of ownership.
So far all I'm getting here has been a knee-jerk statement that I must be a MS weenie. I have used MS software. I have used Linux. I have used Mac OS 6.5/7/8/9. I cut my teeth on an APF Imagination machine. I used a Commodore 64 and a Commodore 128. I am using W2K to write this on.
I am probably more in the average of the technical population than anyone who is a software bigot -- the attitude of only one OS should be out there is troubling. The attitude that their is not room for any one specific OS is disturbing.
A knee-jerk reaction to any product because it is made by MS is on par with a knee-jerk reaction to any student who dresses in a dark trench coat. I expect more of professionals.
Are their any systems designers/planners out there? Anyone who has done any coding on a large scale product? Anyone else who's seen what happens when 90% of the problem is addressed leaving a critical 10% to hang in the wind?
I am not condemning Linux on the desktop -- just pointing out that the product has not been proven to meet general desktop requirements. I would suggest that instead of a knee-jerk reaction of "well you don't know..." you might examine the shortcomings of the product base and honestly attempt to shore it up.
Perhaps I'm asking for maturity in the community.
As part of the maturation of Linux it's got to be out there.
About the limited savings I mentioned -- my point is exactly as stated above -- as you stated it. I have seen no reliable studies by any groups supporting any conjecture that a Linux install is cheaper in any way shape or form. What I have seen is industry standard support costs for Windows environments. It is easy to identify within those industry standard documents the support costs of the licensing of the software. That cost is close to what I represented. I'd suggest you look up Gartner to start with.
If you know of any study by someone with authority please do tell. After all your unsupported rebuttal was pretty much a troll.
Of the thousands of applications available today in vertical markets and of the thousands of applications being developed what percentage are being developed or available on the platform used by 90% of the OS users out there today? Windows is king on the desktop still. A fact that is still a fact no matter how you may try to dispute it (industry surveys of late).
I have worked for state and local agencies and governments in the US only. I can tell you from experience that the majority of programs they have to use and need implemented are windows based. Is this a cart or horse problem? Good question -- the issue/point here is that the bean counters need to make sure they know the answer and most times DO NOT.
The only stupid question in a system integration is the unasked one and most times bean counters need competent IT support to make those decisions. If the bean counters have not heard the reason that the objections I raised are invalid in their configuration then they need to ask more questions.
Mexico City is part of a sovereign nation. If you haven't noticed many nations, Mexico included, have some sort of intellectual property arrangement with the world at large. Intellectual property is protected in some manner.
X and Terminal Server/Citrix server differing but similar purposes. One does not replace the other. Perhaps if you were a little more agnostic you might consider that Windows has a spot in IT today. Terminal Server/Citrix enhances that spot. There is some overlap between X and Terminal Server/Citrix. Overlap that might help to drive competition and enhance the product.
The down side to this is that MS understands licensing. They understand Corporate and Government (national and international) IT. They understand MARKETING. The fact is that a Windows machine can interact in a mixed mode environment at little incremantal cost. A Linux machine will incurr licensing fees to use Windows based applications. It is the way the license is written. Advange: Windows.
I installed NT 4 on 200 workstations 2.5 years ago. We have updated service packs but not re-installed the OS on any machine. I don't get the point of the Redhat comment. Since that time we have installed NT and W2K on hundreds more machines. It is stable. It works today. It runs all the apps the users want to run.
Star Office is a great effort and it is coming of age. Other than having little faith in Sun.... (they are a company.. .they are owned by stockholders and are beholden to those stockholders... if the management is not making a profit they are out on their rears and the new management can orphan or attempt to modify the licensing of the product.) my only issue is that it is a second class citizen in the Office space. It is not the leader, the innovator or the standard. MS Office is.
Frankly, in this troll you have posted, I see no data and no solid argument. Got any facts? You may find I'm very receptive to them.
Because this parity is reality in politics the world over.